[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: Summary of work areas
I'm currently working on a document that classify solutions to the
numerous high-level problems related to multiadressed multihomed IPv6
end-sites. I mention actual propositions (HIP, NOID, SCTP...) only
as examples that illustrate a particular way to solve a problem.
I divided the problems into five categories :
- Destination Locator Retrieval and Selection
- Source Locator Selection
- Failure Detection
- Preservation of Security
- Traffic Engineering
The complete table of content is copy/paste'd below.
The document is not finished yet. I hope it will help fixing the ideas.
-- Cedric
Table of Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 4
2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . 4
3. Destination Locator Retrieval and Selection . . . . . . . . 4
3.1 Retrieval . . . . . . . . . . . . . . . . . . . . . . . . . 4
3.1.1 From the Domain Name System . . . . . . . . . . . . . . . . 4
3.1.2 From a Dedicated Service . . . . . . . . . . . . . . . . . . 4
3.1.3 Using Transport-Level Protocol . . . . . . . . . . . . . . . 4
3.2 Stack Levels for the Destination Locator Selection . . . . . 5
3.2.1 Application-Level . . . . . . . . . . . . . . . . . . . . . 5
3.2.2 Transport-Level . . . . . . . . . . . . . . . . . . . . . . 5
3.2.3 Between IP and Transport Levels . . . . . . . . . . . . . . 5
3.2.4 IP-Level . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.3 Destination Locator Selection Mechanisms . . . . . . . . . . 6
3.3.1 Experimentation-Based Selection . . . . . . . . . . . . . . 6
3.3.2 Using Routing Protocols . . . . . . . . . . . . . . . . . . 6
4. Source Locator Selection . . . . . . . . . . . . . . . . . . 6
4.1 Stack Levels for the Source Locator Selection . . . . . . . 6
4.1.1 Application-Level . . . . . . . . . . . . . . . . . . . . . 7
4.1.2 Transport-Level . . . . . . . . . . . . . . . . . . . . . . 7
4.1.3 Between IP and Transport Levels . . . . . . . . . . . . . . 7
4.1.4 IP-Level . . . . . . . . . . . . . . . . . . . . . . . . . . 7
4.2 Source Locator Selection Mechanisms . . . . . . . . . . . . 7
4.2.1 Using Routing Protocols . . . . . . . . . . . . . . . . . . 7
4.2.2 Automatic Selection by IP Infrastructure . . . . . . . . . . 8
4.2.3 Infrastructure-Driven Selection . . . . . . . . . . . . . . 8
5. Failure Detection . . . . . . . . . . . . . . . . . . . . . 8
5.1 End-to-End Keepalive . . . . . . . . . . . . . . . . . . . . 8
5.2 Passive Detection . . . . . . . . . . . . . . . . . . . . . 9
5.3 Using Routing Protocols . . . . . . . . . . . . . . . . . . 9
6. Preservation of Established Communication Sessions . . . . . 10
6.1 Application-Level . . . . . . . . . . . . . . . . . . . . . 10
6.2 Session-Level . . . . . . . . . . . . . . . . . . . . . . . 10
6.3 Transport-Level . . . . . . . . . . . . . . . . . . . . . . 10
6.4 Between IP and Transport Levels . . . . . . . . . . . . . . 10
6.5 IP Level . . . . . . . . . . . . . . . . . . . . . . . . . . 11
7. Preservation of Security . . . . . . . . . . . . . . . . . . 11
8. Ingress Filtering Issue . . . . . . . . . . . . . . . . . . 11
8.1 Relaxing the Source Address Check . . . . . . . . . . . . . 11
8.2 Source Address Based Routing . . . . . . . . . . . . . . . . 12
8.3 Ensuring Right Source Address Selection by the Host . . . . 12
8.4 Packet Rewriting at Exit Router . . . . . . . . . . . . . . 13
9. Traffic Engineering . . . . . . . . . . . . . . . . . . . . 13
9.1 Outbound Traffic Engineering . . . . . . . . . . . . . . . . 13
9.1.1 Infrastructure-Driven Traffic Engineering . . . . . . . . . 14
9.1.2 Host-Driven Traffic Engineering . . . . . . . . . . . . . . 15
9.2 Inbound Traffic Engineering . . . . . . . . . . . . . . . . 15
Le jeu 13/11/2003 a 05:05, Christian Huitema a ecrit :
> My own list of task includes:
>
> - description of an incremental roadmap that makes "business sense"
> - solving the egress filtering issue (including when addresses cannot be
> rewritten)
> - selection of a first pair of address/locator to "establish contact",
> either from application to TCP (as in the DT2 proposal or in the NOID
> proposal) or from identifier to locator (in the SIM proposal)
> - learning the set of addresses/locators associated to the
> "distinguished address/locator" (common to DT2 proposal and NOID -- the
> DNS is only one of many possibilities)
> - decision algorithm for actually triggering the use of a different set
> of addresses/locators for an ongoing TCP connection (we should consider
> the trade-off between routing events, mobility events, and transport
> events such as retransmit on timer)
> - threat model & possible mitigations of the various attacks
>
> -- Christian Huitema
>
>
> > -----Original Message-----
> > From: owner-multi6@ops.ietf.org [mailto:owner-multi6@ops.ietf.org] On
> > Behalf Of Kurt Erik Lindqvist
> > Sent: Wednesday, November 12, 2003 7:46 PM
> > To: Tony Li
> > Cc: multi6@ops.ietf.org
> > Subject: Re: Summary of work areas
> >
> > -----BEGIN PGP SIGNED MESSAGE-----
> > Hash: SHA1
> >
> >
> > Oh, and I think the bootstrap/start-up face that Margaret and others
> > brought up is something that should be on the list.
> >
> > - - kurtis -
> >
> > On onsdag, nov 12, 2003, at 21:02 Europe/Stockholm, Tony Li wrote:
> >
> > >
> > > Hi,
> > >
> > > To help Elliot out, I'd like for us to start thinking
> > > about our top level work items. As top level items,
> > > they should, IMHO, be as independent as possible (tho
> > > not wholly independent). They should not be nested and
> > > they should not be about the details.
> > >
> > > Here's a strawman:
> > >
> > >
> > > Threat analysis
> > > Locator storage & distribution
> > > Mappings between locators, identifiers, and FQDNs
> > > Security solutions
> > > Exit addressing
> > >
> > > Additions, modifications?
> > >
> > > Tony
> > >
> >
> > -----BEGIN PGP SIGNATURE-----
> > Version: PGP 8.0.2
> >
> > iQA/AwUBP7L+YKarNKXTPFCVEQICiACdFjR3qVo8No4pMrzHdLqJhZ628TUAoLpV
> > CMtZwn3Q2m8V3ibnTY+sIHdn
> > =J+q6
> > -----END PGP SIGNATURE-----
> >